pkg/ansible/controller,pkg/helm/controller; Enable status subresource by default

[dymurray]

Description of the change:
Enable use of status subresource by default in operator-sdk.

Ansible:
Adds support for hybrid managed status between AO and user's Ansible. Get's latest version of resource to ensure that we are updating the latest version of the resource.

Motivation for the change:
In Kubernetes v1.11.0 the status subresource is enabled by default and so this PR enables the status subresource by default in operator-sdk.

Ansible:
Hybrid use case panics when user updates status from Ansible task.

Closes #809

[openshift-ci-robot]

Hi @dymurray. Thanks for your PR.

I'm waiting for a operator-framework member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[shawn-hurley]

@dymurray Before we merge this we need to make sure that we add some migration steps to some sort of doc to describe what it looks like if you're using a CRD without sub-resources and what to do to fix it.

[shawn-hurley]

/hold

this would require v1.12 of kubernetes by default. @fabianvf @dymurray we may need to rethink this, my apolgies I forgot about this caveat when discussing this.

Lets plan on talking at some point today in person to determine what to do.

[dymurray]

Just correcting Shawn's comment, this requires v1.11 of Kubernetes by default:
kubernetes/kubernetes#63598

[hasbro17]

fixes: #809 as we're turning on the status subresource by default which is beta in k8s 1.11 per kubernetes/kubernetes#63598
Although I would like to see a manual confirmation of the status subresource being on by default (and not behind a feature gate) on a k8s 1.11 cluster.

[dymurray]

@hasbro17 I can confirm that bringing up a v1.11.0 Kubernetes Cluster the status subresource is on by default.

[joelanford]

@dymurray Have you looked into the Helm error in this build at all? I'm seeing the same error trying to add a status subresource to a CRD.

https://travis-ci.org/operator-framework/operator-sdk/builds/464000402?utm_source=github_status&utm_medium=notification

[joelanford]

I think it's a bug fixed in 1.11.3 here, but I'm not positive. With 1.11.2 the helm e2e test works if I change the operator-sdk new command to --kind=Memcached. With 1.11.3, the helm e2e test works as is.

[dymurray]

@joelanford I haven't dug into it just yet but your link looks promising. I will report back if I find anything. It only began happening when I added the status writer to the helm controller.

[shawn-hurley]

/hold cancel

[hasbro17]

@dymurray Can you please update the PR title and description to more clearly say that we're enabling the status subresource by default.

[dymurray]

Helm CI tests are related to the bug Joe mentioned above. I ran the tests with kubectl version 1.11.3 and the error went away.

Edit: I bumped the client libraries to 1.11.5 to resolve this and updated Gopkg.* files

[dymurray]

I have removed all client library bumps from this PR because of #807

[shawn-hurley]

LGTM

[hasbro17]

Overall SGTM
@dymurray Can you please include a line in the CHANGELOG under the Changed section to mention that the SDK now generates the CRD with the status subresource enabled by default.
And link to doc/user/client.md doc on how to update the status.

[dymurray]

Thanks @hasbro17, added the note to the changelog.

[hasbro17]

LGTM after nit

[hasbro17]

My bad, I forgot to squash before merging.